Modulation of Insulin Secretion by Fatty Acyl Analogs

Sprague-Dawley male rats (Harlan) weighing 250–350 g were fed a standard powdered diet (55% carbohydrate, 20% protein, 4.5% fat, 12.1% moisture, 3.4% cellulose, and 5% ash). For long-term exposure to M16, animals were fed for 4 weeks with 390 mg M16/kg body wt mixed in the diet. For short-term exposure, animals were gavaged with a single dose of M16 or Cl-DICA (365 mg/kg body wt) in 1% (wt/vol) CMC (carboxy methyl cellulose). Animal care and experimental procedures were in accordance with guidelines of the animal care committee of the Hebrew University.

The tail artery and vein of fed animals were cannulated under local anesthesia using 1.5 ml of 2% lignocaine. Catheter patency was maintained with saline. Following recovery for 60 min, basal plasma insulin, C-peptide, and glucose were determined in two to three blood samples collected before secretagogue infusion. For hyperglycemic clamp, the animal was primed by infusing a glucose bolus (245 mg/kg) followed by maintaining blood glucose at 200 mg/dl for 49 min by variable infusion of 25% glucose in saline through the venous catheter. Blood samples were collected through the arterial catheter, every 2 min during the first 10 min and than every 3–8 min, into heparin tubes and kept on ice. Following 49 min of hyperglycemic clamp, glucose infusion was switched to saline supplemented with 2 ml washed red cells to replace blood loss, and the animals recovered for 60 min. Arginine-stimulated insulin secretion was primed by infusing a bolus of arginine (147 mg/kg body wt) through the venous catheter, followed by constant arginine infusion (4.5 mg/min) for 40 min. Blood samples were initially collected every 2 min during the first 10 min, followed by blood sampling every 5 min. Plasma samples were kept at −20°C for determination of insulin and C-peptide levels.

Animals cannulated as described above were primed by injecting a bolus mixture through the tail vein, consisting of 245 mg/kg body wt glucose and 147 mg/kg body wt arginine, followed by constant infusion of glucose and arginine at a rate of 34 and 18 mg · min⁻¹ · kg⁻¹ body wt, respectively. Following 10 min of constant infusion, C-peptide release was blocked by replacing the glucose-arginine infusion with an infusion of somatostatin (Sigma). The animals were first primed with a bolus of 5 μg/kg body wt somatostatin in 0.6% fatty acid–free BSA in saline, followed by somatostatin infusion at a rate of 0.4 μg · min⁻¹ · kg⁻¹ body wt. Blood samples were collected every 2.5 min, centrifuged, and kept at −20°C for determination of C-peptide levels.

Rats were killed by cervical dislocation and pancreas excised and weighted, followed by an overnight fixation of a tissue sample in Bouin fixative. Pancreases were embedded in paraffin, and 5-μm sections were prepared at three depths separated by 300-μm intervals. The sections were incubated with guinea pig anti-pig insulin (Sigma), followed by goat anti–guinea pig cy5 (Jackson), and were scanned by fluorescent microscopy (Axiovert 100). Images were processed for β-cell area by Image Pro Plus. β-Cell mass was calculated by multiplying the percentage of β-cell area by total pancreatic weight.

Animals were anesthetized with a ketamine (Fort Dodge, IA) xylazine (Biob, France) mixture. Pancreas perfusion was carried out as described by Penhos et al. (31). The perfusate consisted of a modified Krebs-Ringer bicarbonate buffer (pH 7.4) containing 4% dextran T70 (Sigma) and 0.2% BSA (fraction V) and was saturated with 95% O₂/5% CO₂. Following a 20-min recovery with perfusate containing 2.8 mmol/l glucose, samples were collected from the portal vein every 5 min for the next 20 min, followed by switching to 17.0 mmol/l glucose for the next 30 min. Samples were collected every 2.5 min for the first 10 min and then every 5 min. Perfusate samples were centrifuged and the supernatant kept at −20°C for determination of insulin and C-peptide levels. Perfusions were terminated by perfusing the pancreas with Trypan Blue to determine the extent of pancreas perfusion. Only experiments in which the whole pancreas was stained with Trypan Blue were further processed.

Pancreatic islets were isolated as described by Sutton et al. (32) using collagenase P (Roche Diagnostics). Islets were digested and washed in Hank’s solution containing 5.5 mmol/l glucose, handpicked under binocular, and cultured in five batches of five islets in 24-well Nunc plates containing RPMI 1640 (Sigma) supplemented with 100 units/ml penicillin, 0.1 mg/ml streptomycin, and 10% (vol/vol) FCS (Biological Industries) in the presence of 2.8 or 11.2 mmol/l glucose, as indicated, in the absence or presence of 250 μmol/l M16 or Cl-DICA, as indicated. Medium samples were collected as indicated to assess cumulative insulin release. Following 48 h in culture, the islets were washed, preincubated for 60 min in Krebs-Ringer buffer containing 10 mmol/l HEPES, 0.5% fatty acid–free BSA (KRBH-BSA), and 2.8 mmol/l glucose and were subsequently subjected to GSIS, K⁺-stimulated insulin secretion (KSIS), or arginine-stimulated insulin secretion (ArgSIS) in 24-well plates as indicated. For GSIS, the islets were incubated for 45 min in 600 μl KRBH-BSA containing either 2.8 or 20 mmol/l glucose. For KSIS, the islets were incubated for 45 min in 600 μl KRBH-BSA containing 2.8 mmol/l glucose and 30 mmol/l KCl. For ArgSIS, the islets were incubated for 45 min in 600 μl KRBH-BSA containing 2.8 mmol/l glucose and 20 mmol/l arginine. Samples were kept at −20°C for determination of insulin levels. Following incubation, the islets were extracted with 200 μl acidic ethanol (1.8 mol/l HCl in 75% ethanol) and sonicated. The extracts were kept overnight at 4°C, centrifuged, and the supernatant was kept in −20°C for determination of islets insulin content.

Batches of 10 islets cultured for 48 h in 11.2 mmol/l glucose in the presence or absence of 250 μmol/l M16 were incubated in KRBH-BSA containing 0.5 or 20 mmol/l glucose as indicated, followed by extraction with trichloroacetic acid (5% final concentration). The extract was immediately mixed, placed on ice for 15 min, and centrifuged. The supernatant (400 μl) was extracted three times with 1.2 ml diethyl ether, and 0.2 ml of the washed extract were diluted with 0.2 ml buffer containing 20 mmol/l HEPES (pH 7.75) and 3 mmol/l MgCl₂ and kept at −80°C for determination of ATP levels.

Batches of 25 islets were cultured in RPMI medium containing 2.8 or 11.2 mmol/l glucose in the presence or absence of 250 μmol/l M16. Following 48 h in culture, the islets were centrifuged and further incubated for 15 min in KRBH-BSA supplemented with 2.8 or 11.2 mmol/l glucose containing 25 μCi of l-[2,3,4,5-3H]leucine (120 Ci/mmol s.a.) (American Radiolabeled Chemicals, St. Louis, MO). Incorporation of radioactive leucine into proinsulin and total protein was measured as described by Leibowitz et al. (33).

ATP was assayed using a luminometric method (BioLab kit). C-peptide was assayed by radioimmunoassay (RIA) using rat C-peptide kit (Linco, St. Charles, MO). Insulin was assayed by RIA using anti-insulin–coated tubes (ICN Pharmaceuticals, Costa Mesa, CA) and 125I insulin (Linco). Insulin RIA was calibrated by rat insulin.

Effects of short-term exposure to LCFA analogs in vivo were verified in rats administered a single dose of vehicle, M16, or Cl-DICA and subjected to hyperglycemic (11.2 mmol/l glucose) clamp, 2 h following drug administration. Short-term exposure to M16 resulted in an increase in basal plasma C-peptide levels (Table 1). Short-term exposure to M16 and Cl-DICA resulted in twofold activation of insulin and C-peptide secretion under hyperglycemic clamp conditions (Table 1). The ratio of plasma C-peptide to insulin remained unaffected during the clamp, indicating that the observed increases in plasma insulin and C-peptide levels were accounted for by activation of insulin secretion rather than modulation of its plasma clearance. Activation of insulin secretion by M16 or Cl-DICA was glucose specific and not induced when using arginine as the primary secretagogue (Table 1).

Treatment of rats for 4 weeks with M16 did not affect body weight or fasting plasma glucose levels but resulted in nonsignificant decrease in plasma insulin and plasma C-peptide (Table 2). Extended exposure to M16 resulted in a robust decrease in total and first-phase insulin and C-peptide secretion under hyperglycemic (11.2 mmol/l glucose) clamp conditions (Table 2). GSIS suppression by M16 was accompanied by a 28% increase in the ratio of secreted C-peptide to insulin (Table 2), indicating that suppression of plasma insulin levels in response to long-term exposure to M16 was partly accounted for by selective activation of plasma insulin clearance, which is in line with our previous findings in M16-treated fa/fa rats (34). However, plasma C-peptide clearance by renal excretion remained unaffected in treated animals (0.13 ± 0.01 vs. 0.11 ± 0.01 pmol · l⁻¹ · min⁻¹ in nontreated and M16-treated animals, respectively), thus implying genuine suppression of GSIS by chronic exposure to M16. Suppression of insulin secretion by extended exposure to M16 was similarly observed on arginine-induced depolarization (Table 2), implicating the involvement of M16 in a step beyond K_ATP channels in the stimulus-exocytosis coupling pathway.

Suppression of GSIS by long-term exposure to M16 in vivo was further verified under conditions of pancreas perfusion in situ. Basal (3.0 mmol/l glucose) insulin secretion remained unaffected, while GSIS (17.0 mmol/l glucose) decreased by 36% in M16-treated animals (7,756 ± 577 and 4,963 ± 828 ng/30 min [n = 7] in nontreated and M16-treated animals, respectively, P < 0.05 by t test), indicating that suppression of GSIS by long-term exposure to M16 in vivo was not mediated by a washable humoral factor.

Suppression of insulin secretion by chronic exposure to M16 in vivo was accompanied by depletion of islets’ insulin (Table 3). Insulin secretion and islets’ insulin content were suppressed to a similar extent by chronic exposure to M16, thus indicating possible linkage between the two functions. However, β-cell mass, the mean area of β-cell clusters, and the density of β-cell clusters remained unaffected in M16-treated animals (Table 3). Hence, suppression of insulin secretion by chronic exposure to M16 in vivo may imply its involvement in β-cell function rather than β-cell mass.

Suppression of insulin secretion by extended exposure to M16 was further verified in rat islets cultured for 48 h with added M16 (250 μmol/l) in 11.2 mmol/l glucose. Following culturing, the islets were washed in KRBH-BSA containing 2.8 mmol/l glucose, followed by measuring insulin secretion induced by 20 mmol/l glucose (GSIS), 20 μmol/l arginine (in 2.8 mmol/l glucose) (ArgSIS), or 30 mmol/l KCl (in 2.8 mmol/l glucose) (KSIS) in the absence of M16. Basal (2.8 mmol/l glucose) insulin secretion remained essentially unaffected by prior exposure to M16 in the presence of high glucose (11.2 mmol/l). However, GSIS, KSIS, or ArgSIS were robustly suppressed by M16 (Fig. 1A), which is in line with suppression of insulin secretion in vivo by chronic exposure to M16 (Table 2). In contrast to M16, GSIS remained unaffected by exposure of islets to Cl-DICA (Fig. 1B), pointing to a requirement for the respective acyl-CoA. Suppression of insulin secretion by M16 was not accounted for by decrease in islets’ basal or glucose-stimulated ATP content (Fig. 1C).

Suppression of insulin secretion in islets cultured with M16 required the concomitant presence of added glucose (>2.8 mmol/l) during exposure to M16. Thus, insulin secretion in response to glucose remained unaffected by M16 in islets cultured in the presence of basal glucose (2.8 mmol/l) (Fig. 2A). Furthermore, long-term exposure to M16 in basal glucose resulted in M16-stimulated insulin secretion (M16SIS) (Fig. 2A). The mutual relationship between M16 and glucose, namely the requirement for high glucose (11.2 mmol/l) for suppression of insulin secretion by M16 and suppression of M16SIS by high glucose, was further verified by measuring insulin secretion into the culture medium during a 48-h exposure to M16. Exposure to M16 under conditions of basal glucose resulted in robust progressive secretion of insulin to the culture medium, whereas insulin secretion remained unaffected by M16 under conditions of high glucose (Fig. 2B and C). In contrast to M16SIS, insulin secretion by islets cultured in basal glucose remained unaffected by exposure to Cl-DICA (data not shown).

Extended exposure of islets to M16 (250 μmol/l) resulted in moderate decrease in insulin content in islets cultured in low (2.8 mmol/l) glucose (1,082 ± 114 and 808 ± 102 ng/five islets [n = 4] in nontreated and M16-treated islets, respectively, P < 0.01 by paired t test) and in pronounced decrease in islets cultured in high (11.2 mmol/l) glucose (1,208 ± 108 and 625 ± 86 ng/five islets [n = 14] in nontreated and M16-treated islets, respectively, P < 0.0005 by t test). Insulin secretion and islets’ insulin content were suppressed to a similar extent by extended exposure to M16, as reflected by the ratio of insulin secretion (45 min) to islets’ insulin content (4.0 ± 0.02% [n = 7] and 3.7 ± 0.02% [n = 7] in nontreated and in islets exposed to 250 μmol/l M16 for 48 h, respectively), thus indicating possible linkage between the two functions. M16-induced decreases in islets’ insulin content under conditions of basal glucose could essentially be accounted for by the robust increase in insulin secretion prompted by M16 (M16SIS) (Fig. 2). M16-induced decreases in islets’ insulin content under conditions of culturing the islets in high glucose could be ascribed to M16 suppression of glucose-induced insulin biosynthesis, combined with depletion of pancreatic insulin due to sustained hyperglycemia. Indeed, proinsulin biosynthesis was inhibited by 40% in islets cultured with M16 in 11.2 mmol/l glucose but not in islets cultured in 2.8 mmol/l glucose (Fig. 3), thus accounting for the decrease in insulin content induced by M16 under conditions of high glucose. The role played by sustained hyperglycemia in depleting pancreatic insulin was evaluated by culturing the islets in the presence of diazoxide. Suppression of GSIS by M16 in islets cultured for 48 h in high glucose was blunted in the presence of added diazoxide, with concomitant blunted depletion of pancreatic insulin (Fig. 4), indicating that suppression of GSIS by M16 in islets cultured in high glucose was partly accounted for by sustained insulin depletion.

The mode of action of DICA compounds in the β-cell context has been verified here in order to define discrete LCFA metabolites and pathways involved in modulating insulin secretion during short-term or extended exposure to LCFAs. Since DICA compounds are essentially not metabolized, they may highlight LCFA effects, which could be masked by the rapid turnover of natural LCFAs. Furthermore, since Cl-DICA does not serve as a substrate for CoA thioesterification, it may simulate long-chain free fatty acids, while M16 may simulate effects exerted by the free acid, as well as the acyl-CoA thioester of LCFA (29,30). It should be pointed out, however, that quantitative differences between M16 and Cl-DICA could result from respective differences in their intracellular concentrations due to the intracellular entrapment of M16-CoA but not of free Cl-DICA acid.

It is reported here that a single dose of M16 or Cl-DICA results in activation of glucose but not ArgSIS in vivo, thus implicating an early glucose-specific step in activation of the stimulus-exocytosis coupling pathway by LCFAs. Activation of GSIS by short-term exposure to either M16 or Cl-DICA, in spite of the nonavailability of Cl-DICA for CoA thioesterification, may imply that their free-acid form, rather than the respective acyl-CoA, mediates activation of GSIS by DICA or LCFA. The incretin-like amplifying activity of free LCFAs, in the specific context of insulin secretion stimulated by glucose as a primary secretagogue, may specifically implicate GPR40 activation. This proposed mode of action is corroborated by the 1) previously reported activation of GPR40 by M16 (16), 2) specific requirement for GPR40 for the incretin-like activity of LCFAs, as verified in GPR40^−/− mice (35), and 3) specific requirement for the presence of high glucose for GPR40 effects in β-cells (36).

The incretin-like activity of free LCFA transduced by GPR40 may apparently imply that endogenous LCFA-CoAs maintained by glucose-derived malonyl-CoA (37) play a minor role, if any, in promoting GSIS. This is in line with 1) GSIS being unaffected by overexpressed malonyl-CoA decarboxylase (38, but see 39, where overexpression of malonyl-CoA decarboxylase was claimed to suppress GSIS in the presence of added exogenous free fatty acid, but not in their absence), 2) dumping of LCFA-CoA levels by high glucose (40), and 3) Kir6.2 gating by LCFA-CoAs (11,41). It is noteworthy, however, that in light of its glucose specificity, the incretin-like activity of M16 free acid or Cl-DICA (Table 1) may not account for the reported activity of LCFAs in amplifying insulin secretion induced by KCl or arginine (8) or the reported obligatory requirement for LCFAs in KSIS or ArgSIS in fasted rats in vivo (42). Hence, the incretin-like activity of LCFA is proposed to be transduced by two complementary transduction pathways: namely, activation of a glucose-specific early step of insulin secretion transduced by GPR40 and mediated by free LCFAs and activation of a late step in the stimulus-exocytosis sequel, mediated by an LCFA metabolite downstream of acyl-CoA.

In addition to activation of GSIS by short-term exposure to LCFAs or DICA compounds, LCFA and M16 (Fig. 2) activate insulin secretion as primary secretagogues, namely under conditions of basal glucose (6–8). The following differences between the incretin-like and the secretagogue characteristics of LCFAs in the β-cell context are noteworthy. The secretagogue activity is blunted by high glucose (Fig. 2), being replaced by the secretagogue activity of glucose itself. In contrast, the incretin-like activity of M16 prevails under conditions of high glucose (Table 1). The secretagogue activity of amphipathic carboxylates initiated by short-term exposure may proceed on extended exposure (Fig. 2B and C). In contrast, the incretin-like activity is conditioned by short-term exposure, being replaced by suppression of GSIS on extended exposure. Unlike the incretin-like effect, the secretagogue activity is not induced by Cl-DICA, indicating that it is transduced by the acyl-CoA rather than the free-acid form of LCFA. This is corroborated by the previously reported partial inhibition of the secretagogue activity of LCFA by Triacsin C (43). The secretagogue activity of M16 is negligent in vivo compared with glucose (Table 1) and is mostly evident in islets, where it approaches the secretagogue activity of glucose (Fig. 2). In contrast, the incretin-like activity of M16 is of most significance in vivo (Table 1). These differences between the secretagogue and the incretin-like activities of amphipathic carboxylates in the β-cell context may indicate that in spite of their apparent similarity in activating insulin secretion, the two activities are transduced by unique pathways.

Extended exposure to LCFA or M16 results in depletion of insulin stores (5,6), suppressed proinsulin biosynthesis (5,6,44,45), and suppressed insulin secretion in response to glucose (5). Insulin depletion and suppression of insulin secretion were partly alleviated by adding diazoxide (46). Inhibition of insulin secretion by extended exposure to M16 was characterized by 1) suppression of GSIS, KSIS, or ArgSIS, implicating the involvement of a late step in the stimulus-exocytosis coupling pathway, beyond K_ATP channels; 2) requirement for the concomitant presence of high glucose in culture (Figs. 1A and2A); and 3) specificity for M16, whereas Cl-DICA was inactive, thus implicating a requirement for the acyl-CoA form of LCFA.

Suppression of insulin secretion by extended exposure to LCFAs has been previously ascribed to lipotoxicity due to loading β-cells with triglycerides and/or lipotoxic lipids (e.g., ceramides) (24,26,47), to blunted glucose-induced ATP production as a result of LCFA-induced Randle cycle (5), to uncoupling of mitochondrial oxidative phosphorylation (48), or to direct gating of Kir6.2 channels by LCFA-CoA (11). Suppression of insulin secretion by M16 may help in evaluating some of these proposed modes of action. Thus, since M16-CoA is not esterified into lipids, its suppression of insulin secretion may not be ascribed to putative lipotoxic products, which is in line with the preservation of β-cell mass and pancreatic morphology in M16-treated animals and in corroboration with Boucher et al. (49). Similarly, since M16-CoA is not β-oxidized, its suppression of insulin secretion is not accounted for by putative β-oxidation products that may inhibit glucose uptake and its utilization, which is in line with previous reports (5,44,50). Also, maintenance of intracellular ATP levels in M16-treated islets may further refute M16 effects due to uncoupling of oxidative phosphorylation. Lack of suppression by Cl-DICA may indicate that suppression of insulin secretion by LCFA or M16 is mediated by the respective acyl-CoAs, targeting a downstream element in the stimulus-exocytosis coupling pathway that still remains to be defined.

In conclusion, the use of DICA compounds may help in verifying the characteristics and concerned transduction pathways that may transduce the incretin-like secretagogue and the suppressive activities of LCFA in the β-cell context. The incretin-like activity induced by short-term exposure is glucose specific, appears to be mediated by the free acid rather than the acyl-CoA, and is proposed to be transduced by activation of GPR40. The secretagogue activity is blunted by high glucose, maintained during extended exposure, is mostly evident in cultured islets, and transduced by the acyl-CoA. Suppression of glucose, K⁺, or ArgSIS by extended exposure is induced under conditions of high glucose, accompanied by inhibition of glucose-induced proinsulin biosynthesis, and mediated by the acyl-CoA. The respective acyl-CoA target(s) remain to be identified.

We thank N. Kaiser (Hebrew University) for the generous gift of rat insulin and for her helpful advice.